* [dev-test][PATCH 1/2] f2fs: add a missing truncate_inode_pages_final @ 2017-03-10 9:50 Yunlei He 2017-03-10 9:50 ` [PATCH 2/2] f2fs: avoid move encrypted block in BG_GC if it has been in memory Yunlei He 2017-03-17 1:49 ` [dev-test][PATCH 1/2] f2fs: add a missing truncate_inode_pages_final Jaegeuk Kim 0 siblings, 2 replies; 6+ messages in thread From: Yunlei He @ 2017-03-10 9:50 UTC (permalink / raw) To: jaegeuk, yuchao0, linux-f2fs-devel This patch add a missing truncate_inode_pages_final Signed-off-by: Yunlei He <heyunlei@huawei.com> --- fs/f2fs/super.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c index cfb40d3..7ad2606 100644 --- a/fs/f2fs/super.c +++ b/fs/f2fs/super.c @@ -642,6 +642,8 @@ static int f2fs_drop_inode(struct inode *inode) if (f2fs_is_atomic_file(inode)) drop_inmem_pages(inode); + truncate_inode_pages_final(&inode->i_data); + /* should remain fi->extent_tree for writepage */ f2fs_destroy_extent_node(inode); -- 2.10.1 ------------------------------------------------------------------------------ Announcing the Oxford Dictionaries API! The API offers world-renowned dictionary content that is easy and intuitive to access. Sign up for an account today to start using our lexical data to power your apps and projects. Get started today and enter our developer competition. http://sdm.link/oxford ^ permalink raw reply related [flat|nested] 6+ messages in thread
* [PATCH 2/2] f2fs: avoid move encrypted block in BG_GC if it has been in memory 2017-03-10 9:50 [dev-test][PATCH 1/2] f2fs: add a missing truncate_inode_pages_final Yunlei He @ 2017-03-10 9:50 ` Yunlei He 2017-03-13 18:30 ` Jaegeuk Kim 2017-03-17 1:49 ` [dev-test][PATCH 1/2] f2fs: add a missing truncate_inode_pages_final Jaegeuk Kim 1 sibling, 1 reply; 6+ messages in thread From: Yunlei He @ 2017-03-10 9:50 UTC (permalink / raw) To: jaegeuk, yuchao0, linux-f2fs-devel If an encrypted block has been read to memory, we just dirty it and return directly. Signed-off-by: Yunlei He <heyunlei@huawei.com> --- fs/f2fs/gc.c | 11 +++++++++-- 1 file changed, 9 insertions(+), 2 deletions(-) diff --git a/fs/f2fs/gc.c b/fs/f2fs/gc.c index 418fd98..ca94518 100644 --- a/fs/f2fs/gc.c +++ b/fs/f2fs/gc.c @@ -564,7 +564,7 @@ static bool is_alive(struct f2fs_sb_info *sbi, struct f2fs_summary *sum, } static void move_encrypted_block(struct inode *inode, block_t bidx, - unsigned int segno, int off) + int gc_type, unsigned int segno, int off) { struct f2fs_io_info fio = { .sbi = F2FS_I_SB(inode), @@ -607,6 +607,12 @@ static void move_encrypted_block(struct inode *inode, block_t bidx, */ f2fs_wait_on_page_writeback(page, DATA, true); + if (gc_type == BG_GC && PageUptodate(page) && + fscrypt_has_encryption_key(inode)) { + set_page_dirty(page); + goto put_out; + } + get_node_info(fio.sbi, dn.nid, &ni); set_summary(&sum, dn.nid, dn.ofs_in_node, ni.version); @@ -827,7 +833,8 @@ static void gc_data_segment(struct f2fs_sb_info *sbi, struct f2fs_summary *sum, start_bidx = start_bidx_of_node(nofs, inode) + ofs_in_node; if (f2fs_encrypted_inode(inode) && S_ISREG(inode->i_mode)) - move_encrypted_block(inode, start_bidx, segno, off); + move_encrypted_block(inode, start_bidx, + gc_type, segno, off); else move_data_page(inode, start_bidx, gc_type, segno, off); -- 2.10.1 ------------------------------------------------------------------------------ Announcing the Oxford Dictionaries API! The API offers world-renowned dictionary content that is easy and intuitive to access. Sign up for an account today to start using our lexical data to power your apps and projects. Get started today and enter our developer competition. http://sdm.link/oxford ^ permalink raw reply related [flat|nested] 6+ messages in thread
* Re: [PATCH 2/2] f2fs: avoid move encrypted block in BG_GC if it has been in memory 2017-03-10 9:50 ` [PATCH 2/2] f2fs: avoid move encrypted block in BG_GC if it has been in memory Yunlei He @ 2017-03-13 18:30 ` Jaegeuk Kim 2017-03-14 1:37 ` heyunlei 0 siblings, 1 reply; 6+ messages in thread From: Jaegeuk Kim @ 2017-03-13 18:30 UTC (permalink / raw) To: Yunlei He; +Cc: linux-f2fs-devel On 03/10, Yunlei He wrote: > If an encrypted block has been read to memory, we just dirty it > and return directly. > > Signed-off-by: Yunlei He <heyunlei@huawei.com> > --- > fs/f2fs/gc.c | 11 +++++++++-- > 1 file changed, 9 insertions(+), 2 deletions(-) > > diff --git a/fs/f2fs/gc.c b/fs/f2fs/gc.c > index 418fd98..ca94518 100644 > --- a/fs/f2fs/gc.c > +++ b/fs/f2fs/gc.c > @@ -564,7 +564,7 @@ static bool is_alive(struct f2fs_sb_info *sbi, struct f2fs_summary *sum, > } > > static void move_encrypted_block(struct inode *inode, block_t bidx, > - unsigned int segno, int off) > + int gc_type, unsigned int segno, int off) > { > struct f2fs_io_info fio = { > .sbi = F2FS_I_SB(inode), > @@ -607,6 +607,12 @@ static void move_encrypted_block(struct inode *inode, block_t bidx, > */ > f2fs_wait_on_page_writeback(page, DATA, true); > > + if (gc_type == BG_GC && PageUptodate(page) && > + fscrypt_has_encryption_key(inode)) { Why do we need the encryption key? Thanks, > + set_page_dirty(page); > + goto put_out; > + } > + > get_node_info(fio.sbi, dn.nid, &ni); > set_summary(&sum, dn.nid, dn.ofs_in_node, ni.version); > > @@ -827,7 +833,8 @@ static void gc_data_segment(struct f2fs_sb_info *sbi, struct f2fs_summary *sum, > start_bidx = start_bidx_of_node(nofs, inode) > + ofs_in_node; > if (f2fs_encrypted_inode(inode) && S_ISREG(inode->i_mode)) > - move_encrypted_block(inode, start_bidx, segno, off); > + move_encrypted_block(inode, start_bidx, > + gc_type, segno, off); > else > move_data_page(inode, start_bidx, gc_type, segno, off); > > -- > 2.10.1 > > > ------------------------------------------------------------------------------ > Announcing the Oxford Dictionaries API! The API offers world-renowned > dictionary content that is easy and intuitive to access. Sign up for an > account today to start using our lexical data to power your apps and > projects. Get started today and enter our developer competition. > http://sdm.link/oxford > _______________________________________________ > Linux-f2fs-devel mailing list > Linux-f2fs-devel@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel ------------------------------------------------------------------------------ Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot ^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH 2/2] f2fs: avoid move encrypted block in BG_GC if it has been in memory 2017-03-13 18:30 ` Jaegeuk Kim @ 2017-03-14 1:37 ` heyunlei 2017-03-14 3:42 ` Chao Yu 0 siblings, 1 reply; 6+ messages in thread From: heyunlei @ 2017-03-14 1:37 UTC (permalink / raw) To: Jaegeuk Kim; +Cc: linux-f2fs-devel Hi Jaegeuk, On 2017/3/14 2:30, Jaegeuk Kim wrote: > On 03/10, Yunlei He wrote: >> If an encrypted block has been read to memory, we just dirty it >> and return directly. >> >> Signed-off-by: Yunlei He <heyunlei@huawei.com> >> --- >> fs/f2fs/gc.c | 11 +++++++++-- >> 1 file changed, 9 insertions(+), 2 deletions(-) >> >> diff --git a/fs/f2fs/gc.c b/fs/f2fs/gc.c >> index 418fd98..ca94518 100644 >> --- a/fs/f2fs/gc.c >> +++ b/fs/f2fs/gc.c >> @@ -564,7 +564,7 @@ static bool is_alive(struct f2fs_sb_info *sbi, struct f2fs_summary *sum, >> } >> >> static void move_encrypted_block(struct inode *inode, block_t bidx, >> - unsigned int segno, int off) >> + int gc_type, unsigned int segno, int off) >> { >> struct f2fs_io_info fio = { >> .sbi = F2FS_I_SB(inode), >> @@ -607,6 +607,12 @@ static void move_encrypted_block(struct inode *inode, block_t bidx, >> */ >> f2fs_wait_on_page_writeback(page, DATA, true); >> >> + if (gc_type == BG_GC && PageUptodate(page) && >> + fscrypt_has_encryption_key(inode)) { > > Why do we need the encryption key? Here I am afraid some cases dirty encrypted pages have no authority to write back. Thanks. > > Thanks, > >> + set_page_dirty(page); >> + goto put_out; >> + } >> + >> get_node_info(fio.sbi, dn.nid, &ni); >> set_summary(&sum, dn.nid, dn.ofs_in_node, ni.version); >> >> @@ -827,7 +833,8 @@ static void gc_data_segment(struct f2fs_sb_info *sbi, struct f2fs_summary *sum, >> start_bidx = start_bidx_of_node(nofs, inode) >> + ofs_in_node; >> if (f2fs_encrypted_inode(inode) && S_ISREG(inode->i_mode)) >> - move_encrypted_block(inode, start_bidx, segno, off); >> + move_encrypted_block(inode, start_bidx, >> + gc_type, segno, off); >> else >> move_data_page(inode, start_bidx, gc_type, segno, off); >> >> -- >> 2.10.1 >> >> >> ------------------------------------------------------------------------------ >> Announcing the Oxford Dictionaries API! The API offers world-renowned >> dictionary content that is easy and intuitive to access. Sign up for an >> account today to start using our lexical data to power your apps and >> projects. Get started today and enter our developer competition. >> http://sdm.link/oxford >> _______________________________________________ >> Linux-f2fs-devel mailing list >> Linux-f2fs-devel@lists.sourceforge.net >> https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel > > . > ------------------------------------------------------------------------------ Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot ^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH 2/2] f2fs: avoid move encrypted block in BG_GC if it has been in memory 2017-03-14 1:37 ` heyunlei @ 2017-03-14 3:42 ` Chao Yu 0 siblings, 0 replies; 6+ messages in thread From: Chao Yu @ 2017-03-14 3:42 UTC (permalink / raw) To: heyunlei, Jaegeuk Kim; +Cc: linux-f2fs-devel On 2017/3/14 9:37, heyunlei wrote: > > Hi Jaegeuk, > On 2017/3/14 2:30, Jaegeuk Kim wrote: >> On 03/10, Yunlei He wrote: >>> If an encrypted block has been read to memory, we just dirty it >>> and return directly. >>> >>> Signed-off-by: Yunlei He <heyunlei@huawei.com> >>> --- >>> fs/f2fs/gc.c | 11 +++++++++-- >>> 1 file changed, 9 insertions(+), 2 deletions(-) >>> >>> diff --git a/fs/f2fs/gc.c b/fs/f2fs/gc.c >>> index 418fd98..ca94518 100644 >>> --- a/fs/f2fs/gc.c >>> +++ b/fs/f2fs/gc.c >>> @@ -564,7 +564,7 @@ static bool is_alive(struct f2fs_sb_info *sbi, struct f2fs_summary *sum, >>> } >>> >>> static void move_encrypted_block(struct inode *inode, block_t bidx, >>> - unsigned int segno, int off) >>> + int gc_type, unsigned int segno, int off) >>> { >>> struct f2fs_io_info fio = { >>> .sbi = F2FS_I_SB(inode), >>> @@ -607,6 +607,12 @@ static void move_encrypted_block(struct inode *inode, block_t bidx, >>> */ >>> f2fs_wait_on_page_writeback(page, DATA, true); >>> >>> + if (gc_type == BG_GC && PageUptodate(page) && >>> + fscrypt_has_encryption_key(inode)) { >> >> Why do we need the encryption key? > > Here I am afraid some cases dirty encrypted pages have no authority to write back. Not sure this is the right way, I suspect that user may change master key to incorrect one in keyring, then, this approach will make original data in page cache being encrypted with the wrong key and writebacked to storage even if user only have read permission? Thanks, > > Thanks. >> >> Thanks, >> >>> + set_page_dirty(page); >>> + goto put_out; >>> + } >>> + >>> get_node_info(fio.sbi, dn.nid, &ni); >>> set_summary(&sum, dn.nid, dn.ofs_in_node, ni.version); >>> >>> @@ -827,7 +833,8 @@ static void gc_data_segment(struct f2fs_sb_info *sbi, struct f2fs_summary *sum, >>> start_bidx = start_bidx_of_node(nofs, inode) >>> + ofs_in_node; >>> if (f2fs_encrypted_inode(inode) && S_ISREG(inode->i_mode)) >>> - move_encrypted_block(inode, start_bidx, segno, off); >>> + move_encrypted_block(inode, start_bidx, >>> + gc_type, segno, off); >>> else >>> move_data_page(inode, start_bidx, gc_type, segno, off); >>> >>> -- >>> 2.10.1 >>> >>> >>> ------------------------------------------------------------------------------ >>> Announcing the Oxford Dictionaries API! The API offers world-renowned >>> dictionary content that is easy and intuitive to access. Sign up for an >>> account today to start using our lexical data to power your apps and >>> projects. Get started today and enter our developer competition. >>> http://sdm.link/oxford >>> _______________________________________________ >>> Linux-f2fs-devel mailing list >>> Linux-f2fs-devel@lists.sourceforge.net >>> https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel >> >> . >> > > > . > ------------------------------------------------------------------------------ Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot ^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [dev-test][PATCH 1/2] f2fs: add a missing truncate_inode_pages_final 2017-03-10 9:50 [dev-test][PATCH 1/2] f2fs: add a missing truncate_inode_pages_final Yunlei He 2017-03-10 9:50 ` [PATCH 2/2] f2fs: avoid move encrypted block in BG_GC if it has been in memory Yunlei He @ 2017-03-17 1:49 ` Jaegeuk Kim 1 sibling, 0 replies; 6+ messages in thread From: Jaegeuk Kim @ 2017-03-17 1:49 UTC (permalink / raw) To: Yunlei He; +Cc: linux-f2fs-devel Hi Yunlei, I just hit this, so will pop the patch out. [ 870.158785] sysrq: SysRq : Show Blocked State [ 870.158991] task PC stack pid father [ 870.159024] kworker/u8:2 D 0 127 2 0x00000000 [ 870.159033] Workqueue: writeback wb_workfn (flush-8:32) [ 870.159036] Call Trace: [ 870.159042] __schedule+0x28e/0x8d0 [ 870.159045] schedule+0x36/0x80 [ 870.159049] io_schedule+0x16/0x40 [ 870.159052] wait_on_page_bit+0xf7/0x130 [ 870.159054] ? page_cache_tree_insert+0x120/0x120 [ 870.159057] truncate_inode_pages_range+0x4b0/0x930 [ 870.159063] truncate_inode_pages_final+0x4d/0x60 [ 870.159073] f2fs_drop_inode+0x1a0/0x290 [f2fs] [ 870.159075] iput+0xf0/0x230 [ 870.159083] f2fs_sync_inode_meta+0x3f/0x100 [f2fs] [ 870.159090] block_operations+0xfc/0x200 [f2fs] [ 870.159097] ? f2fs_inode_dirtied+0x62/0xd0 [f2fs] [ 870.159106] ? __insert_extent_tree+0xc0/0xf0 [f2fs] [ 870.159113] write_checkpoint+0x9d/0xe50 [f2fs] [ 870.159121] f2fs_gc+0x398/0x480 [f2fs] [ 870.159129] ? write_data_page+0x68/0x90 [f2fs] [ 870.159137] f2fs_balance_fs+0x132/0x150 [f2fs] [ 870.159145] __write_data_page+0x111/0x5c0 [f2fs] [ 870.159153] f2fs_write_cache_pages+0x23a/0x4a0 [f2fs] [ 870.159160] f2fs_write_data_pages+0x9b/0x1d0 [f2fs] [ 870.159162] ? set_page_dirty+0x5b/0xb0 [ 870.159204] ? update_inode+0x273/0x2e0 [f2fs] [ 870.159206] do_writepages+0x1e/0x30 [ 870.159208] __writeback_single_inode+0x45/0x320 [ 870.159211] writeback_sb_inodes+0x266/0x5f0 [ 870.159213] __writeback_inodes_wb+0x92/0xc0 [ 870.159215] wb_writeback+0x268/0x300 [ 870.159217] ? get_nr_dirty_inodes+0x4c/0x70 [ 870.159219] wb_workfn+0x2df/0x410 [ 870.159221] process_one_work+0x1fc/0x4b0 [ 870.159223] worker_thread+0x4b/0x500 [ 870.159224] kthread+0x101/0x140 [ 870.159226] ? process_one_work+0x4b0/0x4b0 [ 870.159227] ? kthread_create_on_node+0x60/0x60 [ 870.159229] ret_from_fork+0x2c/0x40 [ 870.159269] fsstress D 0 3329 3328 0x000000 Thanks, On 03/10, Yunlei He wrote: > This patch add a missing truncate_inode_pages_final > > Signed-off-by: Yunlei He <heyunlei@huawei.com> > --- > fs/f2fs/super.c | 2 ++ > 1 file changed, 2 insertions(+) > > diff --git a/fs/f2fs/super.c b/fs/f2fs/super.c > index cfb40d3..7ad2606 100644 > --- a/fs/f2fs/super.c > +++ b/fs/f2fs/super.c > @@ -642,6 +642,8 @@ static int f2fs_drop_inode(struct inode *inode) > if (f2fs_is_atomic_file(inode)) > drop_inmem_pages(inode); > > + truncate_inode_pages_final(&inode->i_data); > + > /* should remain fi->extent_tree for writepage */ > f2fs_destroy_extent_node(inode); > > -- > 2.10.1 > > > ------------------------------------------------------------------------------ > Announcing the Oxford Dictionaries API! The API offers world-renowned > dictionary content that is easy and intuitive to access. Sign up for an > account today to start using our lexical data to power your apps and > projects. Get started today and enter our developer competition. > http://sdm.link/oxford > _______________________________________________ > Linux-f2fs-devel mailing list > Linux-f2fs-devel@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel ------------------------------------------------------------------------------ Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot ^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2017-03-17 1:50 UTC | newest] Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed) -- links below jump to the message on this page -- 2017-03-10 9:50 [dev-test][PATCH 1/2] f2fs: add a missing truncate_inode_pages_final Yunlei He 2017-03-10 9:50 ` [PATCH 2/2] f2fs: avoid move encrypted block in BG_GC if it has been in memory Yunlei He 2017-03-13 18:30 ` Jaegeuk Kim 2017-03-14 1:37 ` heyunlei 2017-03-14 3:42 ` Chao Yu 2017-03-17 1:49 ` [dev-test][PATCH 1/2] f2fs: add a missing truncate_inode_pages_final Jaegeuk Kim
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.